Non-penalty goals is easily a superior measure of someone's goalscoring than total goals (and non-penalty goals plus assists to then get someone's total scoring contribution).
Penalties are converted at an 80% clip and only one person on a team gets to take them. They inflate tallies purely on the basis that someone is the penalty taker, and could have gone to four or five others on any given team.
If we apply the same logic, we can't count tap-ins cos they're easy (scored at better than an 80% clip), then add the appropriate difficulty modifiers for long range goals/free kicks and one on ones if you're gonna be that finickity about it. Why is penalties the line that you draw?
In fact, we should be discounting all strikers, because playing further forward inflates the chances of a player scoring a goal.
If you're gonna draw arbitrary lines in the sand, just blindly rely on xG ffs.
Ridiculous.
Well, yeah, I would take xG+xA (from a proper model, like StatsBomb or Opta) to be a better measure of talent than NPG+A.
The key idea is that goals scored from the penalty spot don't reflect someone's goalscoring ability in the normal bits of the game, the bits that make up 99.9% of it, and that only one person on each team gets to take penalties (typically/for the most part) so you lop them off to be able to make a proper comparison.